- Design
- retrospective multicohort study with propensity score matching and network meta-analysis
- Population
- 1,062,722 matched pairs hospitalised 2014 to 2018 across 62 US health care organisations, mean age 48.4 years
- Primary outcome
- 14 psychiatric and neurologic diagnoses between 1 month and 2 years after admission
- Effect
- time-lost ratio above 1 in 116 of 140 tests; cognitive deficits 19.08% vs 6.71% after cardiac infection
A propensity-matched study drew on 62 US health care organisations to compare 1,062,722 matched pairs of people hospitalised between 2014 and 2018, mapping 14 psychiatric and neurologic outcomes from one month to two years after admission. Infections of ten different body systems were compared with each other, with non-infectious admissions, and with the general population.
Against non-infectious hospitalisation, the measure of time lost to disease exceeded 1 in 116 of 140 infection-outcome tests. The largest relative excess was encephalitis - for cardiac infections, a median ratio of 6.54 (95% CI 3.78 to 11.31). The largest absolute excess was cognitive deficits: after cardiac infection, 19.08% versus 6.71%, a difference of 12.37 percentage points. Infectious encephalitides were the leading risk factor for 7 of 14 outcomes. Children showed the same direction of association with significantly smaller absolute differences.
The practical value is in the absolute numbers rather than the ratios. A patient discharged after serious sepsis or endocarditis has a materially raised chance of a cognitive or psychiatric diagnosis in the next two years, and it is not routinely looked for. This is not a reason to start a new clinic; it is a reason to ask about cognition and mood at the post-discharge review, and to take a family's report of 'not the same since' as data rather than distress.
- Ask about memory, concentration and mood at review after any hospitalisation for serious infection
- Take collateral history seriously - relatives notice cognitive change before testing does
- Screen rather than assume: a brief cognitive test and a mood screen at 1 to 3 months
- Give the highest index of suspicion after infectious encephalitis, the strongest signal here
- Reserve the same vigilance for delirium during admission, which flags later cognitive risk
Why it matters
It reframes serious infection as a brain-health event with a two-year tail, not an episode that ends at discharge.
Don't overread it
This is observational and matched on recorded characteristics; it shows association, not that the infection caused the later diagnosis.
The statistics, in plain English
The ratio of restricted mean time lost compares how much disease-free time each group loses; above 1 means the infected group loses more. Ratios like 6.54 sound dramatic but apply to rare outcomes such as encephalitis, while the 12.37 percentage-point difference in cognitive deficits is the figure that matters at the bedside because the outcome is common. Propensity matching balances measured characteristics only - severity of illness and pre-existing frailty are imperfectly captured, and they plausibly explain part of this.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for psychiatry, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free